Influenza matrix protein M1

(1)

Membrane Binding Study

D I S S E R T A T I O N

Zur Erlangung des akademischen Titels d o c t o r r e r u m n a t u r a l i u m

(Dr. rer. nat.) im Fach Biophysik

eingereicht an der

Mathematisch-Naturwissenschaftlichen Fakultät I der Humboldt-Universität zu Berlin

von Diplom-Biologin Nadine Jungnick

Präsident der Humboldt-Universität zu Berlin Prof. Dr. Jan-Hendrik Olbertz

Dekan der Mathematisch-Naturwissenschaftlichen Fakultät I Professor Dr. Andreas Herrmann

Gutachter: 1. Prof. Dr. Andreas Herrmann 2. PD Dr. Michael Veit

3. Professor Dr. Thomas Günther Pomorski

eingereicht: 20.9.2011

(2)

Zusammenfassung

Die Aufklärung der Prozesse, die zur Zusammensetzung des Influenza A Virus führen, ist Be- standteil für die Bekämpfung dieser Infektionskrankheit. Der Viruspartikel setzt sich aus einer Hülle, der darunter liegenden Matrix und dem Genom zusammen. Das Genom ist als Bündel aus acht Ribunucleoproteinkomplexen organisiert. Die Hülle besteht aus einer Membran, die mit Sphingomyelin und Cholesterol angereichert ist und den darin eingebetteten Membran- proteinen Hämagglutinin, Neuraminidase und dem Protonenkanal M2. Die unter der Hülle liegende Matrix wird von einem einzigen Influenzaprotein formiert: Dem Matrixprotein M1.

Es spielt eine Schlüsselrolle im Replikationszyklus des Virus in der Zelle. Es interagiert mit dem genetischen Material, mit den Membranproteinen und der Lipidmembran der Hülle.

Die vorliegende Arbeit gibt Auskunft, welche Lipide eine Rolle in der M1-Membran- Wechselwirkung spielen. Die Liste der identifizierten Lipide umfasst neben dem bereits be- kannten Phosphatidylserin auch Phosphatidylglycerol und Phosphatidsäure. Verschiedene Phosphatidylinositole konnten ebenfalls identifiziert werden. Als stärkster M1 Bindungspart- ner trat dabei Phosphatidylinositol-4-Phosphat zutage.

Weitere auf Mutanten basierende Untersuchungen zeigten, dass der membranbindende Be- reich nicht auf eine einzelne Domäne in M1 festgelegt werden kann. Die N-terminale M1- Domäne mit ihrem Oberflächen-exponierten, positiv geladenen Areal und die C-terminale Domäne interagierten mit Modellmembranen.

Das Resultat dieser Interaktionen konnte mittels mikroskopischer Untersuchungen an giganti- schen unilamellaren Vesikeln dokumentiert werden. Für M1 und für eine Mutante, die nur aus der N-terminalen M1-Domäne besteht, konnte eine von anderen viralen Proteinen unabhängi- ge homooligomere Organisation auf der Membran gezeigt werden. Diese M1-Cluster könnten während der Zusammensetzung des Viruspartikels als Fundament für die Eingliederung aller weiteren viralen Komponenten dienen.

Schlagworte: Fluoreszenzmikroskopie, Influenza A Matrixprotein M1, Protein-Membran- Interaktionen, unilamellare Vesikel.

(3)

Abstract

Knowledge about the assembly process of the influenza A virus particle is essential for the development of effective approaches for prevention and treatment of this virus infection. The virus particle consists of an envelope, an underlying matrix, and the encapsulated genome.

The genetic material is organized as bundle of eight ribonucleoprotein complexes that encode for eleven proteins. The envelope consists of a lipid bilayer that is enriched in sphingomyelin and cholesterol. The viral spike proteins, hemagglutinin and neuraminidase, as well as the proton channel M2 are embedded into this membrane. The matrix can be found below the envelope. It is formed by one single protein, the matrix protein M1. M1 plays a crucial role during the replication of the virus in the cell. It interacts with the genetic material, with the envelope proteins and with the lipid bilayer of the envelope.

The results of this study reveal in detail which lipids are targeted by M1. The set of identified lipids contains phosphatylglycerol and phosphatidic acids as new binding partners, beside the known phophatidylserine. Additionally, several phosphatidylinositols were identified. Phos- phatidylinositol-4-phosphate was the strongest binding partner from this group.

Mutant-based analysis revealed that M1 owns more than one membrane binding site. The positively charged area in the N-terminal and the C-terminal domain mediated membrane association of the respective mutant protein.

The final constitution of M1 on the membrane was characterized by confocal fluorescence microscopy on giant unilamellar vesicles. Full length M1 and a mutant that consisted only of the N-terminal part of M1 showed lateral clustering of homooligomers on the vesicle surface.

The clusters formed independently of any other viral component. A function as fundament for the incorporation of the other viral components can be assumed for these clusters.

Keywords: Unilamellar vesicles, influenza matrix protein M1, protein-membrane interaction, fluorescence microscopy

(4)

Abbreviations

APS Ammonium persulfate

Aλ Absorption at wavelength λ in nm

C₆-NBD-PC 1-Palmitoyl-2-[6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl]- sn-glycero-3-phosphatidylcholine

Chol Cholesterol

DLS dynamic light scattering DOPC Dioleoyl-phosphatidylcholine DOPS Dioleoyl-phosphatidylserine EDTA Ethylenediaminetetraacetic acid

FM Fluorescein-5-maleimide

FRAP Fluorescence recovery after photobleaching FRET Förster resonance energy transfer

GUV giant unilamellar vesicle

HA Hemagglutinin

LUV large unilamellar vesicle

M1 Matrix protein 1

MOPS 3-(N-morpholino)propanesulfonic acid NaP 10 mM sodium phosphate buffer, pH 7

NaPKCl 10 mM sodium phosphate buffer with 120 mM potassium chloride, pH 7 Cy-, Cys- Cysteine

N-NBD-DPPE N-(7-nitrobenzy-2-oxa-l,3-di azol-4-yl)-1,2-dipalmitoyl-sn-glycero-3- phosphatidylethanolamine

ODλ optical density at wavelength λ in nm

ORF open reading frame

PAGE polyacrylamide gel electrophoresis

PBS polybasic sequence

PCR Polymerase chain reaction

PI3P Phosphatidylinositol-3-phosphate PI4P Phosphatidylinositol-4-phosphate

SDS Sodium dodecyl sulfate

SM Sphingomyelin

SUV small unilamellar vesicle

Tm melting temperature

TMR 5/6-Carboxy-tetramethylrhodamine-ethyl-maleimide

(5)

1 Introduction

Infection by the influenza virus results in the death of a quarter- to a half-million people an- nually. In Germany alone 21,959 people died in 2008 [1] from flu-like diseases. In April 2009, a highly infectious influenza variant emerged in Mexico and caught world-wide attention.

This influenza A H1N1 strain, so called “New flu”, spread rapidly and was classified as pandemic by the WHO in June 2009 [2]. The fundamental mechanisms governing influenza infection need to be understood in detail to prevent and treat future outbreaks. This effort will require both basic and applied medical research, and studies on the molecular level. Bio- chemical and biophysical methods promise substantial progress in the discovery of the mechanisms of influenza virus replication and assembly in the cell.

1.1 Influenza

The history of the influenza disease can be traced back to the ancient Greeks. Hippocrates described bronchial infections with symptoms similar to influenza [3]. More accurate records of influenza pandemics are available since the 16^th century. During the following centuries several major outbreaks occurred [4] of which the most recent were the Spanish flu with over 50 million deaths in 1918 [5,6] and the Hong Kong flu 1968 with reported one million deaths [7,8].

Wild aquatic birds are the widely accepted reservoir hosts for influenza viruses [9,10]. Re- search on avian influenza-based human viruses revealed a two-step transmission mechanism, where an avian strain infected an intermediate swine host and was then transmitted to humans [11]. Direct transmission between humans and birds has been described for H5N1 virus strains between 2003 – 2005 in Asia [11]. In most cases the transmission rates between ani- mals and humans were low [12]. Pandemic viruses like the H1N1 virus from 2009 were ex- ceptionally contagious and spread rapidly among humans around the world [13,14]. This par- ticular H1N1 virus was first detected in Mexico and characterized on the molecular level. It contained components of human, avian, and swine specific viruses [13,15].

The mechanism upon which different viral genes are combined is called re-assortment and occurs when different influenza hosts live in close proximity and cross infections occur. Dur- ing an infection with different viruses, new virus particles are assembled which contain genetic material from both influenza strains [16].

(8)

The Influenza A and B genus, together with the genera of Influenza C, Thogotoviruses, and the Isavirus belong to the family Orthomyxoviridae. Recently additional viruses have been discovered and pooled in a fifth yet unnamed genus [17] which also harbor the major charac- teristic of the Orthomyxoviridae – the segmented and single-stranded RNA genome of nega- tive polarity. The individual influenza genera can be distinguished by host range, genome and viron structure. Influenza A and B have a similar viron structure and a segmented genome consisting of eight pieces which encode 11 proteins in total (see chapter 1.1.1). Influenza C has in contrast to A and B only seven genomic segments. The distinction between A and B lies in the specificity of the nucleoprotein (NP) which is necessary for the organization of the genome and also determines the host range [18]. Whereas influenza B causes mostly mild symptoms upon infection in humans, pandemics with severe phenotypes are mainly caused by the influenza A virus.

1.1.1 The influenza virus particle

Influenza virus particles are pleomorphic. Beside spheres with an approximate size of 100 – 150 nm [19], filamentous particles can also be observed [20]. The envelope of the particles consists of a lipid bilayer in which three different membrane proteins are embedded (Figure 1). The trimeric hemagglutinin (HA) mediates binding to the host membrane and fusion of the viral and the target membrane. For influenza A 16 different avian HA subtypes (H1 – H16) have been classified by their serological behavior [21,22]. Three HA serotypes (H1, H2, and H3) have been found to be adapted to the human population [23]. The second membrane protein, neuraminidase (NA), is tetrameric and removes the receptor molecules recognized by HA from the outer membrane surface. Nine NA serotypes (N1 – N9) have been described [22]. The tetrameric protein M2, the third membrane protein, acts as a proton channel. Influ- enza B particles contain both HA and NA but M2 is replaced by a proton channel named NB [24]. In Influenza C the multifunctional hemagglutinin-esterase-fusion protein (HEF) carries the activities of HA and NA [18]. The proton channel is formed by the CM2 protein [25].

Directly beneath the envelope lies a shell of M1 protein. M1 oligomerization and interactions with the membrane as well as with HA, NA and M2 have been postulated as being necessary for the formation of this layer [26,27,28]. M1 encloses the virus core that contains the seven (influenza C) or eight (influenza A and B) viral ribonucleoprotein (vRNP) complexes [29]

which carry the genetic material of the virus. Each vRNP complex consists of a single-

(9)

stranded, negatively oriented RNA that is coiled into a panhandle-like structure by oligomer- ized NP. The polymerase complex with the proteins PB1, PB2, and PA is located at the terminal end of the panhandle [30]. The polymerase complex mediates the transcription of viral RNA.

Two additional proteins complete the particle structure. The multifunctional NS1 protein has a modulating effect on the transcription of viral RNA and an enhancing effect on translation of viral mRNA [31]. The nuclear export protein (NEP – previously NS2) is responsible for the export of newly formed vRNP complexes from the nucleus [32].

Figure 1 The influenza particle. (A) Model of the Influenza A particle depicting the envelope with the spike proteins hemagglutinin (HA), neuraminidase (NA) and the proton channel M2 embedded into the lipid bilayer. Oli- gomerized M1 interacts with the envelope components and encloses the viral ribonucleoprotein (vRNP) complexes. Adapted from [33]. (B) Electron micrograph of a spherical and an elongated virus particle with visible spike proteins and genomic material in the center [34]. (C) A model of a vRNP [30].

1.1.2 Influenza replication cycle

The replication of influenza can basically be divided into three stages: (i) internalization, (ii) production of the viral components, and (iii) assembly and release. Figure 2 summarizes the replications cycle. Initially, HA binds to cell surface receptors on the host membrane. These receptors are glycoproteins and glycolipids containing oligosaccharides with terminal sialic acids. Sialic acids comprise a molecule group based on neuraminic acid with the most com- mon member being N-acetylneuraminic acid [35]. The precise linkage of the monosaccha- rides determines the host preference. In case of avian influenza sialic acids are recognized that

(10)

are bound to galactose through α2-3 linkages [23]. Viruses infecting humans attach to α2-6- linked sialic acids [23]. The binding mechanism between HA and the sialic acids carrying glycoproteins and glycolipids is modulated by HAs own oligosaccharides. It was shown for HA with truncated glycosylations that the receptor binding was increased [36,37].

Figure 2 Replication cycle of Influenza A. HA binds in a first step to sialic acids on the surface of the host cell.

The bound virus particle is internalized via receptor-mediated endocytosis. The acidification of the late endosome triggers conformational changes in HA, which lead to fusion of endosomal and viral membrane. The vRNPs are released through the fusion pore and transported to the nucleus. Viral mRNA synthesis occurs in the nucleus. The viral mRNAs induce production of the viral membrane proteins HA, NA, and M2 at the endoplasmic reticulum (ER). These membrane proteins are transported towards the assembly site through the Golgi apparatus. The other proteins are translated on cytoplasmic ribosomes. Early proteins, which are necessary for transcription and vRNP replication and assembly, are imported into the nucleus. Late proteins, like M1 and NS2 (NEP) enable the export of the vRNPs. When all components are assembled at the plasma membrane the new progeny virions bud. Adapted from [38].

When the virus is attached to the host membrane the cell internalizes the particle via receptor- mediated endocytosis. It could be shown that the resulting endosome is produced by clathrin- dependent and -independent pathways [39,40]. The endosome is transported by the cytoskeleton to the interior of the cell in close proximity to the nucleus [41]. Upon formation of the late endosome acidification occurs [42]. The lowering of the pH to values of 5.5 - 5 induces a conformational change in HA. This leads to the exposition and insertion of HA’s hydrophobic

(11)

fusion peptide into the host membrane. In a second rearrangement step of HA the host and viral membrane are drawn together, resulting in fusion and pore formation (Figure 3) [43,44].

Figure 3 Conformational changes in hemagglutinin upon endosome acidification. Upon acidification conformational changes occur in HA (1) that expose the structurally concealed hydrophobic fusion peptide. The fusion peptides of the HA trimer insert into the host membrane (2, 3) and in a second rearrangement step the host membrane is drawn into close proximity to the viral membrane (4). Further action of HA leads to membrane fusion and the formation of a hemifusion diaphragm (5). Finally, a pore is formed (6). Adapted from [43].

Parallel to this, M2 shuttles protons from the endosomal lumen into the viral core. At this point M1 plays a crucial role. It functions as a link connecting the cytoplasmic tails of the three membrane proteins [26] and the virus membranes [27] to the vRNP complexes [45]. The acidification destabilizes the M1 envelope anchor and releases the genome segments into the cytoplasm for transport towards the nucleus [46]. The three nuclear localization signals (NLS) of NP have been shown to be essential for the import of the vRNP complexes into the nucleus.

Import has been shown to be mediated by importin α and β and the nuclear pore complex [41,47,48]. Within the nucleus the trimeric polymerase complex transcribes the negative-sense viral RNA segments into mRNA and replicates them via complementary RNA intermediates into abundant copies of vRNA [49,50]. Simultaneously, the mRNAs are exported to the cytoplasm or endoplasmatic reticulum (ER) where translation of the eleven influenza proteins occurs. HA, NA and M2 are synthesized on the ER from where they are transported via the Golgi apparatus to the apical plasma membrane [51]. A co-localization of HA and NA with cholesterol- and sphingomyelin-enriched microdomains, so called membrane rafts, was shown [52,53,54]; whereas M2 was proposed to be peripherally associated to these liquid ordered membrane domains since only a minor portion of M2 was found to be associated with deter-

(12)

gent-resistant membranes [55]. The newly synthesized cytoplasmic proteins M1, NS1, NP, and the polymerase subunits are targeted to the nucleus by their intrinsic NLS. In case of NEP transport may occur by diffusion [32]. In the cytoplasm the remaining M1 was found to be colocalized with cellular membranes including the Golgi vesicles [56,57,58] independent of other viral proteins. The PB1-F2 protein, a splicing variant of the polymerase subunit PB1, interacts with mitochondrial proteins [59]. In the nucleus, PB1, PB2, PA, and NP together with vRNA are assembled into new vRNP complexes. Their export is triggered by a signal cascade initiated by the accumulation of HA at the plasma membrane [60]. The single vRNP complexes are exported from the nucleus after stepwise association of proteins [61]. First, M1 attaches to the vRNPs. This association inactivates the polymerase complex and arrests it at the terminal part of the vRNP complex [62,63]. NEP binds to M1. It contains a nuclear export signal sequence and is the target for the host protein Crm1 [64]. Crm1 facilitates together with RanGTP the export of vRNPs through the nuclear pore [65]. The re-import of the complexes into the nucleus is inhibited by M1 which possibly shields the NLS of NP [46,66]. Once the M1 covered vRNPs are released into the cytoplasm, transport towards the apical plasma membrane is mediated by cytoskeleton filaments [62,67]. Assembly of all components into the viral particle occurs at the apical plasma membrane, since only there vRNPs are present.

The current model states that M2 and raft associated HA and NA are linked to each other by polymerized M1 that is bound to the cytoplasmic tails of the membrane proteins and to the membrane. The observation of the M1 oligomerization process should be experimentally possible but has not yet been accomplished. Interaction of M1 with membrane proteins and M1 self-oligomerization [68,69,70] are well studied events, and M1 oligomerization is considered essential for membrane bending and formation of the bud involving the cortical actin filaments [71]. The vRNPs are delivered to the budding site and internalized into the bud by pull- ing into the cavity by envelope-M1 and M1-vRNP interactions and additionally by pushing through actin filaments [62]. The mechanism how the single vRNPs are organized in the cytoplasm is based on RNA-RNA interaction of specific signal sequences leading to an oc- tameric super-vRNP [72,73,74]. This complex could be visualized [29] in virus cross-sections.

The incorporation of the eight vRNPs is not the critical but an essential part of particle formation. Budding of empty virus-like particles could also be induced in the absence of vRNPs [75,76]. The last step is the closure of the bud. Here M2 plays a role in organizing the membrane before pinching-off of the new particle [34,55]. Release of the new particle requires

(13)

removal of sialic acids from the cell surface. Otherwise the newly synthesized particles re- main bound on the infected cell. This cleavage is accomplished by NA. The catalytic site of NA binds specifically to the sialic acid of the glycoproteins and glycolipids and hydrolyzes the linkages to the respective terminal sialic acid [77,78].

1.1.3 M1 as key organizer at the membrane level

Mutational studies showed an essential function of M1 during virus particle formation. Infec- tious virus production failed when M1 was lacking or when specific amino acid sequences of M1 were substituted [79]. An additional hint for the importance of M1 arose when the concentration of available M1 was limited. This leads to a delayed and reduced budding of viral particles [80]. Even though its role in virus maturation is of such importance, the function of M1 during the replication cycle is not understood in detail.

The 252 amino acids containing matrix protein is encoded on segment seven, the second smallest of the eight genome pieces. M1 can be isolated from the virus as a 27.8 kDa protein.

Its isoelectric point is 9.81. Several attempts have been undertaken to crystallize the protein.

Sha and Luo [81] presented in 1997 the structure of the N-terminal domain (AA 2 – 158) isolated from virus at pH 4. This structure of wild type M1 revealed an N-terminal subdomain formed by the first four helices (H1 – H4). A flexible link including the short helix H5 con- nects the N-terminal subdomain to the middle subdomain that consists of four additional helices. The C-terminal structure could not be solved due to proteolysis [81]. Subsequently two structures of the same N-terminal part of M1 were published and showed only minor folding differences to the structure published by Sha and Luo. Harris et al. as well as Arzt et al. de- termined the structure at neutral pH [68,82]. Furthermore, the structure of a M1 mutant with substituted NLS signal [83] was solved and revealed no structural differences to the wild type.

Based on the crystal structures, the surface potential of this part of M1 was calculated and local concentrations of either positive or negative charged amino acids could be visualized.

The polybasic sequence (PBS) in H6 that contains the three residues (R101, K104, R105) of the NLS motif (101–RKLKR–105) together with K95 and K98 forms a prominent positively charged surface area. Beside the function of the NLS as a nuclear sorting signal a function as membrane binding site was proposed and electrostatic interactions to negatively charged membrane surfaces were suggested as the binding mechanism [27]. Experimental data to prove this mechanism are so far only available to a limited extent. When purified M1 was

(14)

exposed to liposomes and liposomes with bound M1 were separated from unbound M1, protein association was only detected for negatively charged vesicles. The negatively charged vesicles in this artificial system were produced from a mixture of specific lipids, namely phosphatidylcholine, cholesterol, and negatively charged phophatidylserine [27,84]. However, interactions with other lipids are not reported. Membrane suspensions derived from lysed cells were also tested. Here a M1 association was visualized [27]. Since cellular membranes consist of more than three different lipids, the role of the huge lipid spectrum needs to be further specified. It is possible that phosphatidylserine is not the only M1 target.

Figure 4 3D structure of the N-terminal part of M1. (A) Superimposed ribbon cartoon of the crystal structures at pH 4 (gold) and pH 7 (blue). The major secondary structure elements are nine α-helices linked by loop regions.

β-sheets are not visible. Adapted from [68]. (B) Electrochemical surface potential of the protein from two different views. The protein exhibits enrichment of negative charges (red) and a prominent positively charged (blue) patch including the PBS (95–KAVKLYRKLKR–105) which is located on H6. Modified from [83]

When M1 was expressed in mammalian epithelial cells it was found in the nucleus and attached to membranes [56]. The plasma membrane, the membranes of Golgi apparatus and ER as well as transport and storage vesicles were targeted by M1. Studies performed with fluorescent proteins showed a spotted distribution of M1 [58]. Experiments with fluorescent ER and Golgi markers revealed a co-localization of M1 with the Golgi apparatus [85]. Why M1 was found in association with the Golgi is matter of ongoing research.

(15)

1.2 Lipid membranes

Lipid membranes play a crucial role in both the structure and the function of all – prokaryotic and eukaryotic – cells. They define the cellular space in total (plasma membrane) as well as in eukaryotic cells internal organelles (mitochondria and plastids) and compartments (nuclear membrane, smooth and rough endoplasmic reticulum, Golgi apparatus, endosomes, and ly- sosomes).

They display a uniform overall structure, basically a six nanometer thick double layer of lipid molecules with embedded (integral) and peripheral proteins. Membranes function as barriers which enable regulated transport of molecules from outside inwards and vice versa by protein activity. In addition, many cellular enzymes are attached to membranes where they are concentrated on a lateral plane and facilitate efficient interactions [86]. A continuous exchange of vesicles maintains transport of cargo which is not directly transported via the cytoskeleton or diffusion through the cytoplasm. Beside this, membranes serve as an assembly site for cellular proteins, like clathrin coated pits, or for virus particles.

1.2.1 Cellular lipids

Mammalian cells contain more than thousand different lipid species [87]. Lipids fulfill three general functions: (i) energy storage, (ii) matrix of cellular membranes, (iii) first and second messengers in signal transductions and molecular recognition processes.

Glycerophospholipids are the major lipid component in the membrane. These lipids are com- posed of three components: (i) the headgroup molecule which is attached via a phosphoester bond to (ii) glycerol. Two fatty acid chains (iii) can be found at the other originally hydroxy- lated C-atoms of the glycerol, also attached via ester bonds. Major glycerophospholipids found in the cell are phosphatidic acids (PA), phosphatidylcholines (PC), phosphatidyletha- nolamines (PE), phosphatidylserines (PS), and phosphatidylinsitols (PI) with one to three additional phosphorylations [88].

The second abundant group of molecules is the group of sphingolipids. These molecules show a trimodular composition similar to glycerophospholipids. The skeletal structure component is sphingosine. The amino group of this molecule functions as attachment site for fatty acids, the resulting molecule is classified as ceramide. Substituents can be bound to the primary hydroxyl group of sphingosine through esterification. The most prominent sphingolipid sphingomyelin is depicted in Figure 5. Glycoglycerolipids and glycosphingolipids are glycosylated

(16)

forms of membrane lipids [86]. The carbohydrate moiety of the headgroup can range from a single sugar to very complex polymers.

Figure 5 Membrane lipids of epithelial Madin-Darby canine kidney cells. Cholesterol, phosphatidylcholine, and phosphatidylethanolamine represent the major components of epithelial cells. Less abundant are PI, PS, SM, diacylglycerol, PA, PG, ceramides, and gangliosides like GM3. Adapted from [88].

The attached fatty acids differ in chain length and number of double bonds within the chain.

The degree of unsaturation in the side chains and the concentration of lipids with unsaturated side chains influence the membrane fluidity. The more unsaturated side chains are present the higher the fluidity of the membrane is.

Another group of membrane molecules are sterols including cholesterol, the most abundant sterol in mammalian cells. These molecules incorporate between phospholipids and sphingo-

(17)

lipids. Hydrogen bonds between the sterols hydroxyl group and the lipid headgroups as well as hydrophobic interactions are the driving force for this incorporation.

Minor components of membranes are free fatty acids, lysophospholipds, mono- and diacyl- glycerides, as well as polyisoprenoid lipids.

The lipid composition is dynamic. For epithelial morphogenesis a major shift of the composition from sphingomyelin to glycosphingolipids, together with an increase in phosphatidylethanolamine and cholesterol content could be observed whereas the opposite changes took place during an epithelial-to-mesenchymal transition [88].

1.2.2 Cellular lipid transport and membrane traffic

Glycerophospholipids are mainly synthesized at the interface of cytosol and membrane. The bacterial phospholipid synthesis is located at the cytosolic site of the plasma membrane, whereas eukaryotic cells produce their lipids mainly at the smooth endoplasmic reticulum.

The synthesized lipids are incorporated into the cytoplasmic site of the membrane. The trans- location of lipids from one leaflet to the other is accomplished by flip-flop mechanisms. Flip- flop can occur spontaneously or controlled by membrane proteins called flippases. These proteins bind lipids on one membrane leaflet, transport in an energy-dependent step and release the bound lipid at the other leaflet (reviewed in [89]). Action of P-type ATPases like yeast Dnf1p and Dnf2p allows the maintenance of the different compositions of inner and outer leaflet of the plasma membrane. They keep the content of aminophospholipids like PS and PE in the outer leaflet low by specific transport to the inner leaflet where PS is enriched [90].

The ER is the major lipid production site in eukaryotic cells. It overlaps with the nuclear membrane. The other compartments achieve their supply of lipids by two transport mechanisms. Lipid transfer proteins shuttle specific lipids between membranes whereas vesicles allow transport of formed membranes. Transport with lipid carrying proteins involves a con- tinually growing group of proteins. These proteins act at membrane contacts sites between the ER and the Golgi apparatus or the Golgi and the plasma membrane. The ceramide transporter CERT was active between ER and Golgi membranes [91]. Nir2 was identified as a PI/PC transfer protein, shuttling PC from the Golgi to the ER and phosphatidylinositol back, from which phosphatidylinositol-4-phosphate (PI4P) is subsequently synthesized at the Golgi membranes [92]. The activity of these proteins enriches specific lipids at the respective compartments.

(18)

The major transport form for lipids is the vesicle. A huge and still growing number of proteins are involved in the establishment and trafficking of vesicles in the cell, regulating vesicle identity, local lipid synthesis, and vesicle targeting. Clathrin-coated vesicles can be formed from the plasma membrane as well as the Golgi apparatus. They mediate endocytic events and endosomal vesicular traffic [93,94,95]. COP (coat protein) vesicles of type I are involved in bidirectional transport in the Golgi apparatus as well as recycling of proteins from the Golgi to the ER [96]. Vesicles organized by COPII emerge from the ER and export secretory proteins towards the Golgi complex [97,98].

All these mechanisms taken together enable the cell to sort lipids and proteins specifically to distinct sites of the cell. These transport pathways are used by viruses, like influenza, to sort their components. For example, influenza hemagglutinin highjacks the secretory pathway on its way from the ER through the Golgi to the viral assembly site at the plasma membrane [99].

1.2.3 The plasma membrane

As mentioned before, membranes function as barriers. This function is especially important at the outermost membrane, the plasma membrane. Nutrients and liquids are taken up through this membrane. Signals are perceived and transmitted inwards. Beside these major functions of the regular cellular life cycle other events can be observed here. Assembly and budding of the new influenza viral particles occur at the plasma membrane [62].

The lipids provide the matrix for incorporated proteins. The plasma membrane exhibits a pro- nounced difference in the lipid composition of the cytoplasmic and extracellular leaflets. Hu- man erythrocytes were extensively analyzed as a model system for the plasma membrane.

Their outer leaflet contains mostly phosphatidylcholine and sphingomyelin while the inner leaflet consists mainly of phosphatidylethanolamine and phosphatidylserine. The phosphati- dylinositolphosphates are also not homogeneously distributed across the plasma membrane.

Phosphatidic acid, phosphatidylinositol and phosphatidylinositol-4,5-bisphosphate were mainly present at the inner leaflet [100]. This assymetric appearance of lipids leads to a negatively charged cytoplasmic lipid leaflet (see Figure 6A). Beside this leaflet dimorphism a lateral heterogeneity of the lipid distribution was found in the plasma membrane. Lipid domains called rafts are enriched in cholesterol, sphingomyelin, and glycerophospholipids which carry saturated fatty acids, and specific proteins [101,102]. The local enrichment of these specific

(19)

lipids creates a liquid ordered phase state. These liquid ordered plasma membrane rafts are surrounded by lipids in a liquid disordered organization [103].

Clustering of proteins was observed in a sterol-dependent manner [104]. The incorporated proteins carry specific raft signals, like cysteine palmitoylation close to the transmembrane domain [105] or glycosylphosphatidylinositol (GPI) anchors with saturated fatty acids [106].

Interestingly, the influenza hemagglutinin was found among those proteins, since it clusters cholesterol dependent in mammalian cells [52,53]. There are hints that rafts are also present at intra cellular membranes like the ER [107] and they may serve as sorting platforms in the cell.

Figure 6 summarizes the transversal lipid distribution and presents a model of the lateral organization of the plasma membrane.

Figure 6 Composition and organization of the plasma membrane. (A) Leaflet composition of the plasma membrane. Adaptad from [100]. SM and PC were shown to reside mostly in the outer leaflet; PS and PE were detected in the inner leaflet. Phosphoinositols and its phosphorylated derivatives were distributed on both leaflets.

(B) Model of the lateral heterogeneity within the plasma membrane. GPI-anchored proteins, transmembrane raft proteins cluster together with glycosphingolipids and cholesterol into rafts. Non-raft proteins are excluded.

GSL = Glycosphingolipid, GPL = Glycerophospholipid. Modified from [108].

1.2.4 Influenza protein mediated plasma membrane modification

The influenza proteins HA, NA, and M2 belong to the group of integral membrane proteins.

All three proteins own one α-helical membrane domain which consists mainly of hydrophobic amino acids [109,110,111]. A lipid raft association was shown for HA and NA [109,112]. The intrinsic equipment that facilitates HA raft association was identified. Mutation of specific amino acids in the transmembrane domain of HA revealed the amino acids 530-WILWISFAI-538 as essential for HA raft association [109]. The transmembrane domain and cytoplasmic tail of HA together own three cysteine residues (C551, C559, C562 in HA of the H7N1 strain) which are acylated [113]. The raft association of HA was reduced when

(20)

these cysteines were substituted [53,114]. The raft targeting signals of NA were also assessed via mutagenesis of specific amino acids in the transmembrane domain and cytoplasmic tail.

The mutation of the amino acids 27-GINIISIWIS-35 in NAs transmembrane domain led to reduced raft association of NA [110]. Acylation of NA was not reported. From this data it was surmised that HA and NA recruit raft lipids to the assembly site of the virus particle (reviewed in [51,62,115]). The enrichment of HA and NA in rafts at the budding site is the first step during influenza assembly. But the final particle formation can only be accomplished when the membrane is curved, a bud is developed, and fission of the elongated bud occurs. Virus-like particles (VLPs) were formed, when HA or NA were expressed alone in cells [75,116]. Analy- sis of HA in model membranes revealed, that full length HA or its transmembrane peptide do not bent or tubulate the membrane when they are reconstituted into giant unilamellar vesicles (GUVs) [117]. Similar results were obtained when HA was analyzed in giant plasma membrane vesicles that resemble the plasma membrane. Therefore, cellular factors must be involved in the HA based VLP production. Combined expression of HA, NA, M2 and M1 en- hanced VLP formation significantly [75]. This could have been the result from more efficient bud closure and fission of the closed VLPs. An impact of the influenza proton channel M2 in fission of virus particles was shown by production of mutant virus particles with altered M2.

The mutation did not hinder the bud formation, but the release of the formed viruses was stopped [111,118]. Bending of membranes independent of other influenza proteins was shown when M2 was reconstituted into GUVs. The 17 amino acids long amphipathic helix in the cytoplasmic tail of M2 mediated this bending [111]. M2 was not found to be a raft protein, even though it carries a palmitoylation in its cytoplasmic tail [115,119]. Co-localization of M2 with HA was shown [120,121]. An interaction between HA and M2 would link M2 to the other viral spike proteins and could induce M2 mediated membrane bending.

A role for M1 in the orchestration of the budding process has been proposed. Several studies showed a direct influence of M1 on the shape of the final virus [80,122]. Single mutations in the NLS region of M1 could induce a shift from spherical viruses towards filamentous ones [123]. The change of the viral shape needs modification of the viral membrane and membrane-bound proteins. M1 was not able to induced VLP formation when it was expressed alone in cells. The VLP production was triggered efficiently when HA, NA, M2, and M1 were expressed together. The formation of a M1 layer inside of the VLPs could be shown [75]. This indicated a connection between these four proteins at the membrane level. An interaction of

(21)

M1 with the cytoplasmic tails of all three membrane proteins was shown [28,57,58].

Interestingly, M1-M2 interactions influence the shape of the budding influenza particle [28,120,124]. A current model by Rossman et al. [115] summarizes all the described protein- membrane and protein-protein interactions (Figure 7). The model considers clustering of HA and NA in lipid raft domains as an initial step that slightly bends the membrane at the assembly site. This bud initiation is further accomplished by M1 binding to the cytoplasmic tails of HA and NA and M1-mediated vRNP-attachment to the spike proteins (Figure 7A).

The next step in the virus formation is elongation of the bud by polymerization of M1. This leads to a perpendicular encapsulation of the vRNPs in relation to the plasma membrane. M2 is recruited to the periphery of the budding site through interactions with M1 (Figure 7B). The amphipathic helix of M2 alters membrane curvature at the neck of the bud when it is inserted into the raft boundary. This leads to release of the budding virus (Figure 7C and D).

Figure 7 Model for influenza virus budding at the plasma membrane. (A) Initiation of the virus bud by clustering of HA (red) and NA (orange) in lipid raft domains. M1 (purple) binds to the cytoplasmic tails of HA and NA and cross-links the vRNPs (yellow) to the spike proteins. (B) Elongation of the bud by polymerization of M1. This leads to a perpendicular orientation of the vRNPs to the membrane. M2 (blue) is recruited to the periphery of the budding site through interactions with M1 and/or HA. (C) Insertion of the M2 amphipathic helix at the raft phase boundary alters membrane curvature at the neck of the bud and leads to release of the budding virus. (D) Over- view of budding influenza viruses seen on the membrane. The sorting of HA and NA into lipid rafts (yellow), the formation of a filamentous virion, and membrane scission caused by M2, are depicted. Taken from [115].

(22)

A function of M1 as key organizer in the assembly of influenza virus can be derived from the M1 controlled formation of spherical or filamentous virions and its interaction with the three viral membrane proteins and the vRNPs. Even though the implications of the M1-HA/NA/M2 interaction for lateral lipid sorting and membrane shape changes have been analyzed to a great extent (reviewed in [51,62,115] it is still elusive how M1s intrinsic membrane binding capa- bility fits into this model. Therefore it is essential to know what M1 does when it is attached to the membrane. GUVs and other artificial liposomes can provide a good starting point for examination of M1 membrane association. For instance, the appearance of HA and M2 was successfully visualized on this kind of vesicles [111,117].

1.2.5 Artificial liposomes as model membrane systems for protein interaction

Different membrane model systems have been established to approach protein-membrane interactions at the molecular level. The simplest are artificial liposomes. Such liposomes can be produced in different sizes by various methods.

Small unilamellar vesicles with a diameter below 100 nm were produced by ultrasonification of preformed multilamellar vesicles [125]. They can be used as drug delivery systems [126].

Large unilamellar vesicles (LUVs) were produced from multilamellar vesicles by extrusion through polycarbon membranes with defined pore sizes from 100 nm up to 400 nm [127,128].

Soluble molecules can be encapsulated into the lumen of the LUVs and later released in a controlled way [129]. LUVs cannot only harbor molecules; they can also provide a surface for attachment. ArfGAP1, a GTPase–activating protein (GAP) for ADP-ribosylation factor 1 (Arf1), could interact with lipid membranes. The curvature of the LUVs played a significant role, since binding of ArfGAP1 was stronger when smaller LUVs with a higher membrane curvature were used [130]. The membrane composition was shown to play a significant role for protein-membrane interaction beside the membrane curvature. Positively charged proteins bound to LUVs with negatively charged lipids through electrostatic interaction. This was shown for influenza M1. M1 was mixed with LUVs made of phosphatidylcholine, cholesterol, and negatively charged phosphatidylserine. This mixture was then adjusted to high density with a high concentrated sucrose solution. Sucrose solutions of lower concentrations were overlaid to form a gradient with the LUV-M1 mix at the bottom. Centrifugation was performed. M1 could be detected in the fractions from the low density region of the gradient [84]. It was not detected in the low density fractions when 500 mM NaCl, an inhibitor for

(23)

electrostatic interactions [131], was added to the initial LUV-M1 mix [27]. Thus, flotation with LUVs is a convenient method to analyze the parameters (like certain lipids as binding partners) that influence M1’s membrane interaction.

GUVs are suitable for microscopic methods. They can be produced via induced electro- swelling from surface-attached preformed membranes [132,133]. These micrometer sized vesicles allow direct visualization of protein binding to membranes under the chosen experimental conditions. Suitable dyes for labeling both proteins and membranes are required for this purpose. During the last decades a plethora of fluorescent lipid analogues has been developed for membrane labeling [134]. Usually, they can be added directly to the mixture of the desired lipids in low concentrations (less than 1% of the total amount of lipids) and give well detectable signals [135]. Two strategies can be applied for protein labeling. The protein could be visualized via antibodies which carry a dye [136] or the dye is attached to the protein. A short tetracysteine stretch could be genetically engineered to the protein of interest. These tetracysteine motif was then specifically labeled by biarsenic dyes [137].

Another protein labeling method employs a chemical reaction between a reactive amino acid, for example lysine or cysteine, and the dye. A widely-used reaction couple is tetramethylrhodamine-6-maleimide and a cysteine of the protein [138].

(24)

2 Aims of this study

If M1 is a key organizer in the influenza assembly, it has to fulfill several functions. One of them is the binding to membranes. It is still elusive how this happens and whether specific lipids play a role in this process. Binding of M1 to liposomal membranes of different lipid compositions was investigated to address this question. The flotation assays as well as confocal fluorescence microscopy were chosen as methods for this analysis. Beside the affinity test to validate potential lipid specificity, structural information needs to be recorded. Nothing is known about the intraprotein dynamics upon membrane binding. CD spectroscopy offers the possibility for monitoring molecular rearrangements in presence of membrane material.

Secondary structure and membrane association of genetically engineered mutants were analyzed and compared with the wild type M1 to characterize the domain structure of M1 and domain specific function. Essentially, two questions can be answered by using such mutants.

First, deletion mutants may unravel the location of membrane binding site(s) in M1. Second, site directed mutations in the PBS permit the elucidation of whether this sequence plays a significant role at the molecular level of the membrane interaction of the N-terminal M1 domain. Furthermore, it is possible to validate the molecular rearrangements of these mutants in comparison to the wild type. It may be possible to locate where molecular interactions take place. For instance, when rearrangements could only be detected in a deletion mutant lacking the N-terminal part but not in a mutant without the C-terminal part, the C-terminus of M1 could be identified as putative membrane interaction site of M1.

It is possible to study membrane association and lateral dynamics of M1 in real time by using GUVs of distinct composition as target membranes. Standard and confocal fluorescence mi- croscopes and respective techniques, e.g. fluorescence recovery after photobleaching provide the technical and methodological equipment to monitor a possible self-oligomerization of the protein on membranes. Indeed, an M1-M1 interaction was proposed as the driving force for enrichment of all viral components at the assembly site in the plasma membrane.

Deeper insights into M1’s function during the assembly of the influenza particle can be achieved when the lipid specificity, the membrane binding site and the lateral organization of M1 on the membrane are elucidated.

(25)

3 Materials and Methods

3.1 Instruments

AMINCO-Bowman Series 2 – Luminescence spectrometer (Thermo – Fisher Scientific, Schwerte, Germany)

Biophotometer plus (Eppendorf, Hamburg, Germany) CD Spectrometer J-720 (Jasco, Gross-Umstadt, Germany)

Centrifuge Avanti J-20XP (Rotors JA 25.50 and JLA 10.500) (Beckmann Coulter, Krefeld, Germany)

Confocal microscope IX81 with FluoView-1000 scan head (Olympus, Tokyo, Japan) Fluorescence microscope X100 (Olympus, Tokyo, Japan)

FluoroMax-4 (Horiba Yobin Yvon, Unterhaching, Germany) FluoStar Optima (BMG Labtechnologies, Offenburg, Germany) FUJIFILM FLA-3000 (Fujifilm, Düsseldorf, Germany)

Osmometer type 6 (Löser Messtechnik, Berlin, Germany)

Semi-Dry transfer cell “TransBlot SD” (Bio-Rad, Munich, Germany) Thermal Cycler “MyCycler” (Bio-Rad, Munich, Germany)

Ultracentrifuge Optima L-100K (Rotors 45Ti, 70.1Ti, SW40Ti, SW60) (Beckmann Coulter, Krefeld, Germany)

Ultracentrifuge TL-100 (Beckmann Coulter, Krefeld, Germany)

UV-Vis Spectrometer Lambda 40 (Perkin Elmer Instruments, Waltham, USA) Zetazizer Nano (Malvern, Herrenberg, Germany)

(26)

3.2 Materials

3.2.1 Enzymes, antibodies, kits, and other “ready-to-use” tools

Calf intestine phosphatase, T4 DNA ligase, Restrictionenzymes: DpnI, NdeI, XhoI (Fermen- tas, St. Leon-Rot, Germany)

Taq DNA polymerase (supplied with 10x PCR buffer) (Qiagen, Hilden, Germany or Peqlab, Erlangen, Germany)

Phusion™ DNA Polymerase Kit (supplied with 5x HF PCR buffer) (Finnzymes, Espoo, Finland)

QIAprep Spin Miniprep Kit, QIAquick Gel Extraction Kit (Qiagen, Hilden, Germany)

IgG from goat anti M1, IgG from goat anti H3N2 (Virostat, Portland (Ma), USA)

IgG from donkey anti goat, horseradish peroxidase conjugated (Santa Cruz Biotechnology, Heidelberg, Germany)

ECL™-Kit = Amersham ECL™ Advanced Western Blotting Detection Kit (GE Healthcare Life Science, München, Germany)

Micro BCA Protein Assay Kit (PIERCE, Rockford, USA)

Apoptest™ -FITC Kit, #A700 (VPS Diagnostics, Hoeven, Netherlands)

PIP Strips™ (Echelon Biosciences, Salt Lake City , USA)

(27)

3.2.2 Plasmids and Oligonucleotides

The plasmid pHH21-vM [139] was kindly provided by the group of PD Dr. Michael Veit (Free University, Berlin, Germany). pHH21-vM contains the viral genome fragment for M1 and M2. pET15b (Novagen – Merck, Darmstadt, Germany) was used as acceptor for all cloning procedures.

Figure 8 Map of the plasmid pET15b. The plasmid provides ampicillin resistance. It contains a lac operator for IPTG inducible protein expression behind the T7 promoter. The multiple cloning site includes a thrombin restriction site and the coding sequence for an His-tag. Open reading frames of proteins for recombinant expression were cloned between the restriction sites of NdeI and XhoI. The serine prior the NdeI restriction site was mutated to cysteine for CM1 labeling with fluorophors.

(28)

All PCR oligonucleotides listed below are shown in 5’ – 3’ orientation and were ordered from Invitrogen (Karlsruhe, Germany) in desalted purity. The final oligonucleotide concentration in the PCR mixture was 0.4 µM.

M1-NdeI-fw1 GGGAATTCCATATGAGTCTTCTAACCGAGGTTG (NdeI)

M1-XhoI-rev1 CCGCTCGAGTCACTTGAATCGTTGC (XhoI)

M1-fw GACCACAGAGGTGGCATTTG

M1-rev TCCCATCCGTTTCTGGTAGG

NP-fw AAGAGCCCTTGTGCGTACTG

NP-rev GCTCGTTGTGCTGCTGTTTG

M1-1-fw GCCGTCAAACTATACGCGGCGTTGGCAGCTGAGATAACATTCTATGG

M1-1-rev CCATAGAATGTTATCTCAGCTGCCAACGCCGCGTATAGTTTGACGGC

M1-2-fw CCAAATAACATGGATGCAGCCGTCGCACTATACGCGGCGTTGGCAGC

M1-2-rev GCTGCCAACGCCGCGTATAGTGCGACGGCTGCATCCATGTTATTTGG

NM1rev1 CCGCTCGAGTCACTGTCTGTGAGACCGATGC (XhoI)

CM1-for GGGAATTCCATATGGTGGCTACCACCAATCC (NdeI)

M1Cy GGTGCCGCGCGGCAGCCATATGTGTCTTCTAACCG

M1Cy rev CGGTTAGAAGACACATATGGCTGCCGCGCGGCACC

CM1Cys GGTGCCGCGCGGCTGCCATATGGTGGCTACCACC

CM1Cys–rev GGTGGTAGCCACCATATGGCAGCCGCGCGGCACC

T7-promotor TAATACGACTCACTATAGGG

T7-terminator GCTAGTTATTGCTCAGCGG

(29)

3.2.3 Bacteria and culture media

DH5α (E. coli) F-, endA1, recA1, hsdR17 (rk- mk+), supE44, λ-, thi-1, gyrA(Na1), relA1, Φ80 lacZ∆M15∆ (lacZY A-argF)

Rosetta (E. coli) F–, ompT, hsdSB (rB–, mB–), dcm, gal, lacY1, λ (DE3), pLysS (Cm^R) BL 21 (E. coli) F–, ompT, hsdSB (rB–, mB–), dcm, gal, λ(DE3)

LB + Amp 1 % [w/v] Bacto™ Tryptone; 0,5 % [w/v] Bacto™ Yeast Extract; 0,5 % [w/v] NaCl; 50 µg/ml Ampicillin in ddH2O

LB + Amp + Cm LB + Amp + 50 µg/ml Chloramphenicol Agar plates for

DH5α

1 % [w/v] Bacto™ Tryptone; 0,5 % [w/v] Bacto™ Yeast Extract; 0,5 % [w/v] NaCl; 1,5 % [w/v] Agar; 50 µg/ml Ampicillin in ddH2O

Agar plates for Rosetta

1 % [w/v] Bacto™ Tryptone; 0,5 % [w/v] Bacto™ Yeast Extract; 0,5 % [w/v] NaCl; 1,5 % [w/v] Bacto Agar™; 50 µg/ml Ampicillin; 50 µg/ml Chloramphenicol in ddH2O

LB+G 1 % [w/v] Bacto™ Tryptone; 0,5 % [w/v] Bacto™ Yeast Extract; 0,5 % [w/v] NaCl; 50 µg/ml Ampicillin; 50 µg/ml Chloramphenicol; 0,4 % [w/v]

Glucose in ddH2O

Bacto™ Tryptone, Bacto™ Yeast Extract, and Bacto Agar™ were ordered from BD (Becton, Dickinson and Company, Heidelberg, Germany). Ampicillin and Chloramphenicol were purchased from Sigma Aldrich (Taufkirchen, Germany) and Serva (Heidelberg, Germany). So- dium chloride was of analytical reagent grade. Glycerol for freezing of bacteria was delivered by Roth (Karlsruhe, Germany).

3.2.4 Buffers

All used chemicals in buffers were purchased in analytical reagent grade.

3.2.4.1 Buffers for protein purification

Lyses buffer 50 mM sodium phosphate; 250 mM NaCl; 10 mM EDTA, 20 mM DTT;

pH 7.0 + 16 µg/ml DNase I; 300 µg/ml lysozyme; 1 mM PMSF was added shortly before use

(30)

Washing buffer 1 mg/ml desoxycholate; 20 mM DTT; 1 mM EDTA, pH 7.0; 0.2 mg/ml lysozyme was added directly before use

Unfolding buffer 100 mM sodium phosphate; 1 mM EDTA; 6 M guanidine hydrochloride;

50 mM reduced glutathione; pH 7.0

Refolding buffer 100 mM sodium phosphate; 1 mM EDTA; 0.5 mM oxidized glutathione;

pH 7.5

Binding buffer 10 mM sodium phosphate, 120 potassium chloride, 20 mM imidazole, pH 7 Elution buffer 10 mM sodium phosphate, 120 mM potassium chloride, 250 - 500 mM imi-

dazole; pH 7.0

NaP buffer 10 mM sodium phosphate; pH 7.0

NaPKCl buffer 10 mM sodium phosphate; 120 mM KCl; pH 7.0 Mops buffer 10 mM MOPS; pH 7.0 (KOH)

3.2.4.2 Buffers for SDS-PAGE, Coomassie-staining, and silver staining 4x non-reducing

sample buffer

5 % [w/v] SDS; 0.05 % [w/v] bromine phenol blue; 25 % [v/v] glycerol;

12.5 % [v/v] 1M Tris/HCl buffer pH 6.8

4x reducing sample buffer

= 4x non-reducing sample buffer + 25 % [v/v] β-mercaptoethanol

1x running buffer 192 mM glycine; 25 mM Tris; 3.5 mM SDS

Table 1 Composition of stacking and separating gel

Components for 2 SDS-gels 5 % Stacking gel 12 % Stacking gel 15 % Stacking gel

ddH2O 1.7 ml 3.3 ml 2.3 ml

30 % Acrylamide / Bisacryla- mide (“Rotiphorese Gel 30”

Roth, Karlsruhe, Germany)

0.5 ml 4.0 ml 5.0 ml

0.5 M Tris/HCl; pH 6.8 0.75 ml

1.5 M Tris/HCl; pH 8.8 2.5 ml 2.5 ml

10 % SDS [w/v] 30 µl 100 µl 100 µl

10 % APS 30 µl 100 µl 100 µl

TEMED 3 µl 4 µl 4 µl

(31)

Coomassie staining solution

0.25 % [w/v] Coomassie Brilliant Blue R – 250, 45 % [v/v] ethanol, 10 % [v/v] acetic acid

washing solution 40 % [v/v] Ethanol, 7.5 % [v/v] acetic acid or 50 % [v/v] methanol; 10 % [v/v] acetic acid

Fixation solution 30 % [v/v] ethanol, 10 % [v/v] acetic acid Cross-linking solu-

tion

30 % [v/v] ethanol, 0.5 % [v/v] glutaraldehyde, 0.2 % [w/v] sodium thiosul- fate, 0.5 M sodium acetate

Silver solution 0.1 % [w/v] silver nitrate, 0.02 % formaldehyde Developer 2.5 % [w/v] sodium carbonate, 0.01 % formaldehyde Stop solution 0.05 M EDTA

3.2.4.3 Materials for preparation of large unilamellar vesicles (LUV)

The lipids 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1,2-dioleoyl-sn-glycero-3- phospho-L-serine (DOPS), 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE), 1,2- dioleoyl-sn-glycero-3-phosphate (DOPA), L-α-phosphatidylinositol-4-phosphate (PI4P), N- palmitoyl-D-erythro-sphingosylphosphorylcholine (SM), 1,2-dipalmitoyl-sn-glycero-3- phosphoethanolamine-N-(7-nitro-2-1,3-benzoxadiazol-4-yl) (N-NBD-DPPE), 1-oleoyl-2-6- [(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl-sn-glycero-3-phosphocholine (C6-NBD- PC), 1-oleoyl-2-6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl-sn-glycero-3-phos- phoserine (C6-NBD-PS), and 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-(lissamine rhodamine B sulfonyl) (LR-DOPE) were from Avanti Polar Lipids (Alabaster (Al), USA).

Cholesterol (Chol), Dipalmitoyl-L-α-phosphatidylinositol-3-phosphate (PI4P), and Dipalmi- toyl-L-α-phosphatidylinositol-4-phosphate (PI3P) were from Sigma Aldrich (Taufkirchen, Germany). 1,2-dioleoyl-sn-glycero-3-phosphoglycerol (DOPG) was ordered from Fluka (Buchs, Switzerland). The LUVs were produced in NaP, NaPKCl, NaP with 400 mM KCl or Mops buffer.

3.2.4.4 Flotation assay buffers

25 % Sucrose buffer 25 % [w/v] Sucrose in NaP or NaPKCl or NaP buffer with 400 mM KCl 75 % Sucrose buffer 75 % [w/v] Sucrose in NaP or NaPKCl or NaP buffer with 400 mM KCl

(32)

3.2.4.5 Buffers for PIP™ strips

TBST or TBS 10 mM Tris, 150 mM NaCl, pH 8.0, 0.1 % [v/v] Tween-20 TBST+BSA or

TBS+BSA

10 mM Tris, 150 mM NaCl, pH 8.0, 3 % [w/v] fatty acid free BSA (Sigma- Aldrich) - (0.1 % [v/v] Tween-20) in TBST+BSA

3.2.4.6 Solutions for labeling of proteins

FM solution 2 mM fluorescein-5-maleimide (Invitrogen) in NaPKCl

TMR solution 0.5 mM 5/6-carboxy-tetramethylrhodamine ethyl-maleimide (emp Biotech GmbH, Berlin, Germany) in DMSO

3.2.4.7 Materials and buffers for GUV preparation and microscopy

The lipids DOPC, DOPS, 1,2-dioleoyl-sn-glycero-3-phosphate (DOPA), N-NBD-DPPE, C6- NBD-PC, C6-NBD-PS, Chol, PI4P, and PI3P were from Sigma Aldrich. 1,2-dioleoyl-sn- glycero-3-phosphoglycerol (DOPG) was ordered from Fluka (Buchs, Switzerland). The GUVs were produced in swelling buffer.

Swelling was carried out either on titanium slides or on indium tin oxide coated glass slides.

Swelling buffer 250 mM sucrose + 15 mM NaN3, adjusted to 280 mosm

Microscopy buffer 5.8 mM NaH2PO4, 5.8 mM Na2HPO4, 250 mM glucose (300 mosm) 1x binding buffer 10 mM Hepes, 150 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2

(part of the APOPTEST™-FITC Kit - VPS Diagnostics, Doeven, NL)

3.3 Methods

3.3.1 Cloning of M1

The M1 open reading frame was amplified in a standard PCR with Phusion DNA polymerase from the pHH21vM [139] with the oligonucleotides M1-NdeI-fw1 and M1-XhoI-rev1. These oligonucleotides provided the restrictions sites for NdeI and XhoI in 3’ and 5’ overhanging sequences. The reaction condition in the 20 µl sample were 30 s at 98 °C initial denaturation, 30 cycles with 10 s denaturation at 98 °C followed by 45 s at a oligonucleotide melting tem-

(33)

perature (Tm) of 59 °C and 30 s at 72 °C extension, final extension step for 10 min at 72 °C and cooling to 4 °C until the sample was removed from the cycler. The fragments were doubly digested directly after amplification (5 u NdeI in buffer O for 2 h at 37 °C and after addition of 10 u XhoI another 2 h at 37 °C, volume 20 µl). The acceptor vector pET15b was digested in parallel and dephosphorylated. Purification of the fragments and vector occurred after gel electrophoresis using the QIAquick Gel Extraction Kit (Qiagen, Hilden, Germany). The M1 fragments were inserted into pET15b using T4 DNA ligase according to the manufacturer’s manual. The resulting plasmid was named pET15b-M1. Five µl of the ligation were transfected into 50 µl chemical competent DH5α cells and they were cultivated on ampicillin plates. A part from one growing clone was added to a 25 µl Taq polymerase containing colony PCR mixture with the oligonucleotides M1-fw and M1-rev for M1 sequence detection and amplified (15 min at 95 °C, 25 cycles with 30 s at 95 °C + 45 s at T_m = 50 °C + 1 min at 72

°C, 10 min at 72 °C and cooling to 4 °C). Bacteria clones that contained the sequences of interest were cultured overnight and plasmids were purified with the QIAprep Spin Miniprep Kit (Qiagen, Hilden, Germany) and sent for sequencing (SMB – Services in Molecular Biol- ogy, Berlin, Germany). All plasmids were sequenced between the T7 promoter and T7 terminator region of pET15b using the T7 promoter + T7 terminator oligonucleotides. For long term storage 30 % glycerol [v/v] was mixed into the bacteria culture and frozen at -80 °C.

3.3.2 Cloning of the mutants M1m, NM1, NM1m, and CM1

For site-directed mutagenesis of the polybasic sequence (PBS) in M1 a double overlap extension PCR [140] approach was performed with Phusion polymerase. The principle of this method is based on a complementary oligonucleotide couple which carries the mutation (M1-1-fw + M1-1-rev, M1-2-fw + M1-2-rev) and two the fragment defining primers (M1-NdeI-fw1 + M1-XhoI-rev1).

The first fragment amplification from pET15-M1 occurred in three parts. A PCR with M1- NdeI-fw1 together with M1-1-rev amplified the 5’ part of the fragment (Tm = 59 °C). In a parallel reaction the 3’ part was amplified with M1-1-fw and M1-XhoI-rev1. The two fragments were analyzed electrophoretically, purified, mixed in equal volumes of the purified solution and used in a third subsequent PCR reaction containing only M1-NdeI-fw1 + M1-XhoI-rev1.

The resulting product functioned as matrix for amplification of the second mutagenesis step.

This time M1-NdeI-fw1 + M1-2-rev and M1-2-fw + M1-XhoI-rev1were used. The final PCR

(34)

product was digested with NdeI and XhoI as mentioned before, purified and ligated into pET15b. The resulting plasmid was named pET15b-M1m, transfected and checked via colony PCR with M1-fw and M1-rev (T_m = 50 °C).

The sequences of the N-terminal deletion mutants NM1 and NM1m with mutated PBS were amplified with M1-NdeI-fw1 and NM1rev1 from the plasmids pET15b-M1 for NM1 and pET15bM1m for NM1m (Tm = 59 °C) and cloned into pET15b giving the plasmids pET15b- NM1 and pET15b-NM1m.

The oligonucleotides for the C-terminal deletion mutant CM1 coded on the final plasmid pET15b-CM1 were CM1-for and M1-XhoI-rev1 and used as described above.

3.3.3 Cloning of an additional cysteine into M1, M1m, NM1, NM1m, and CM1

For incorporation of an additional cysteine via S2C substitution into the open reading frames for M1, M1m, NM1, and NM1m, the oligonucleotides listed below were used in a quick change PCR reaction. This method was adapted from a procedure developed by Stratagene [141]. The plasmids pET15b-M1, pET15b-M1m, pET15b-NM1, and pET15b-NM1m functioned as matrices. These plasmids were amplified in a single PCR with the complementary mutagenesis oligonucleotides M1Cy and M1Cy rev. CM1 has no cysteine in its open reading frame. The serine located two amino acids ahead in the N-terminal overhang was therefore converted to cysteine in the plasmid pET15b-CM1 with CM1Cys and CM1Cys-rev. The plasmids with S→C conversion were named pET15b-M1-Cy, pET15b-M1m-Cy, pET15b-NM1- Cy, pET15b-NM1m-Cy and pET15b-CM1-Cys. Oligonucleotides, which did not define a fragment as it is the case in standard PCR, were the basis for this method. Here they in fact function as the origin for undetermined amplification with Phusion polymerase in both direc- tions (30 s at 98 °C; 16 cycles with 10 s at 98 °C, 15 s at 69 °C and 3 min at 72 °C; 10min at 72 °C and cooling to 4 °C). To prevent transformation of non-mutagenized plasmids, the original methylated plasmid DNA was digested with DpnI (45 min, 37 °C, 1x buffer tango) and the remaining plasmids were transformed into DH5α cells. Selected clones were analyzed by colony PCR. The purified plasmids were sequenced, and stored at -20 °C

3.3.4 Expression of M1 and its mutants

For expression of the cloned proteins the respective plasmids were transformed into chemical competent E. coli BL21 or Rosetta cells. Several conditions were tested for the expression